Leadership and organizational behaviour · Module 3

Motivation, incentives and talent

What do incentives really do to effort, honesty and who stays, and how should an organization find, grow and keep the people it most needs?

Beyond the IB course UndergraduateMBAProfessional About 17 minutes · 2,934 words · checked 16 November 2026

From the IB course to this module

The IB course covers the motivation theories of Taylor, Maslow and Herzberg and, at HL, McClelland, self-determination, equity and expectancy theory, with financial and non-financial rewards, recruitment, appraisal and labour turnover (2.4), set within human resource planning (2.1). This module adds the economics of incentive design, the evidence on when pay helps and when it backfires, and talent management as a strategic system.

Section 01: Conceptual foundation

An incentive changes what a person expects to gain from an action, and it works through two channels. The incentive effect changes how hard, and on what, people work. The sorting effect changes who joins and who stays: a scheme that pays for output attracts people who expect to produce a lot and repels those who do not. Many of the largest effects of pay come through sorting, which is easy to miss because it changes the workforce rather than any one person's behaviour.

The central difficulty is measurement. An organization can pay only for what it can observe, and what it can observe is rarely the whole contribution it wants. Whenever a measure is incomplete or can be manipulated, stronger incentives bring more distortion as well as more effort. Kerr (1975) called this the folly of rewarding A while hoping for B.

Motivation is not only extrinsic. People work for interest, mastery, pride and a sense of doing something worthwhile, and incentives interact with those motives, sometimes reinforcing them and sometimes crowding them out.

Talent management applies the same thinking to the supply of capable people over time: which capabilities to build inside and which to buy in the market, how far to differentiate between people, and how to keep those whose value is hardest to replace.

Section 02: Advanced explanation

The economics of incentives: effort, risk and measurement

Agency theory pictures the employer as a principal who cannot observe the effort of an employee, the agent, only an outcome that depends on effort and on luck. Paying for the outcome motivates effort but loads risk onto the employee, who must be compensated for bearing it. In the standard linear model the best commission rate on a measure is b = 1 ÷ (1 + r × c × σ²), where r is the employee's risk aversion, c the cost of effort and σ² the variance of the noise in the measure.

A worked example. With r × c = 0.5 and σ² = 2, the best rate is 1 ÷ (1 + 1) = 0.5: the employee keeps half of each unit of measured output. If the measure becomes three times as noisy (σ² = 6), the rate falls to 1 ÷ (1 + 3) = 0.25. Noisy measures call for weak incentives.

Two design principles follow. The informativeness principle (Holmström, 1979) says pay should use any measure that adds information about effort, including colleagues' results: relative performance evaluation filters out shocks that hit everyone, such as a recession. Its cost is that comparing colleagues invites sabotage and collusion. The controllability idea says people should not be paid on what they cannot influence, though removing everything outside an employee's control can also remove the incentive to cope with it.

The evidence that incentives change behaviour is strong. When Safelite, an American installer of car windscreens, replaced hourly pay with piece rates, Lazear (2000) found that productivity rose by about 44%, roughly half because existing workers produced more and half through sorting, as the firm attracted and kept more productive installers.

Multitasking, targets and gaming

Holmström and Milgrom (1991) showed why the simple model often fails inside organizations. Most jobs contain several tasks, and some of them, such as quality, advice, cooperation and care for the firm's reputation, are far harder to measure than others. Paying strongly for the measurable tasks pulls effort away from the rest. The best response may be weaker incentives than a single-task model suggests, or a redesign of jobs so that hard-to-measure tasks are not bundled with easy-to-measure ones.

Non-linear schemes add timing games. Oyer (1998) found that sales in American firms bunched around financial year ends, consistent with salespeople pulling deals forward or pushing them back to clear annual quota thresholds. Thresholds, accelerators and competitions create cliff edges where a small extra sale earns a large reward, which is exactly where the temptation to cut corners is greatest.

Goals work in a similar way. Goal-setting theory rests on hundreds of studies showing that specific, difficult goals raise performance more than an instruction to do one's best (Locke and Latham, 2002). Ordóñez and colleagues (2009) argued that the benefits had been oversold: narrow goals lead people to neglect what is not measured, encourage risk-taking and dishonesty, and can corrode culture. Locke and Latham (2009) replied that the critics generalized from a few vivid cases. The fair reading is that goals are powerful, which is precisely why their side effects need designing for.

When pay crowds out motivation

Psychologists and economists long disagreed about whether rewards undermine intrinsic motivation. Deci, Koestner and Ryan's (1999) meta-analysis of experiments found that tangible rewards that are expected and depend on doing or completing an interesting task reduce people's later free-choice interest in it, while verbal praise tends to raise it. An earlier meta-analysis (Cameron and Pierce, 1994) had found little overall harm; much of the dispute turned on how studies were grouped.

Field evidence shows the effect is not confined to the laboratory. When Israeli day-care centres introduced a small fine for parents who collected their children late, late collections rose, and stayed high after the fine was withdrawn (Gneezy and Rustichini, 2000a): a moral obligation had become a price. In other experiments a small payment produced worse performance than no payment at all, while a larger one restored it (Gneezy and Rustichini, 2000b). Frey and Oberholzer-Gee (1997) found that offering Swiss residents compensation reduced their stated willingness to host a nuclear waste repository, turning a civic duty into a transaction.

Bénabou and Tirole (2003) gave an economic explanation: a reward is also a signal, that the task is unpleasant or that the principal does not trust the agent. The practical synthesis is that pay works strongly for routine, measurable work; for interesting, professional or prosocial work, how a reward is framed, and whether it signals distrust, can matter as much as its size.

Talent management: make, buy and differentiate

Cappelli (2008) criticized two failing approaches to talent. Long-range succession planning assumed a stable demand for managers that rarely materialized, and buying talent whenever needed proved expensive and imported people without firm-specific knowledge. Cappelli proposed borrowing from supply chain management: forecast needs over shorter horizons, balance make and buy, develop people in smaller steps, and run internal markets so that employees can move to where they are needed.

The war for talent thesis (Michaels, Handfield-Jones and Axelrod, 2001) urged firms to invest disproportionately in their best performers and remove the weakest. Pfeffer (2001) argued that this demoralized the majority and neglected the systems that make ordinary people effective. The evidence on stars supports some caution: Groysberg, Lee and Nanda (2008) found that star equity analysts who moved to another bank performed worse for years, because much of their performance depended on colleagues and systems they left behind. Talent is less portable than its price suggests.

Promotion can be read as a tournament (Lazear and Rosen, 1981): the prize for winning motivates everyone competing for it, so pay gaps between levels raise effort below. Tournaments also reward sabotage and discourage cooperation. A simulation by Scullen, Bergey and Aiman-Smith (2005) suggested that forced ranking, in which a fixed share of staff must receive the lowest rating, yields its largest gains in the first years and diminishing ones after. Ratings of potential, the basis of most succession plans, are subjective and deserve particular scrutiny for bias.

Section 03: Models and frameworks

Where each comes from, how to use it, and where it stops being useful.

Vroom (1964)

Expectancy theory

Motivation depends on three beliefs multiplied together: that effort will lead to performance (expectancy), that performance will lead to a reward (instrumentality) and that the reward is valued (valence). If any one is close to zero, motivation collapses whatever the others.

How to use it

Diagnose a scheme by asking which link is weakest. Do employees believe the target is reachable? Do they trust that hitting it will be rewarded? Do they want what is offered? Most failed schemes break at instrumentality, where trust in management is low.

Where it fails

It assumes people calculate, which fits large, deliberate choices better than daily effort. The beliefs are hard to measure, and the theory is silent on intrinsic motives and on how rewards change the meaning of the work.

Victor H. Vroom (1964). Work and Motivation. John Wiley & Sons.

Holmström (1979); Holmström and Milgrom (1991)

The principal-agent model with multiple tasks

Pay for a measure should rise with how informative and controllable the measure is and fall with its noise and the employee's risk aversion. When a job has tasks of differing measurability, strong pay on the measurable tasks diverts effort from the others.

How to use it

List every task the job really involves and how well each is measured. Where important tasks are poorly measured, weaken individual output pay, use broader measures, move the tasks to a separate role, or rely on supervision and culture instead.

Where it fails

The model assumes people respond only to money and that the employer knows the noise and risk parameters, which it rarely does. It also leaves out fairness, social comparison and intrinsic motivation.

Bengt Holmström (1979). Moral hazard and observability. Bell Journal of Economics, 10(1). Bengt Holmström and Paul Milgrom (1991). Multitask principal-agent analyses: incentive contracts, asset ownership, and job design. Journal of Law, Economics, and Organization, 7 (special issue).

Deci and Ryan (1985)

Self-determination theory

People are most self-motivated when three needs are met: autonomy, competence and relatedness. Rewards that are experienced as controlling undermine intrinsic motivation; rewards experienced as information about competence can support it.

How to use it

Before adding an incentive, ask whether it will feel like control or recognition. Design work so that people choose how to reach goals, receive feedback on their competence and work within teams they value.

Where it fails

Much of the evidence comes from experiments with students on short, interesting tasks; effects in long-term employment are harder to establish. The theory says less about routine work, where extrinsic pay clearly works.

Edward L. Deci and Richard M. Ryan (1985). Intrinsic Motivation and Self-Determination in Human Behavior. Plenum Press. Edward L. Deci, Richard Koestner and Richard M. Ryan (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6).

Section 04: Real business applications

Real organizations appear only with dated, cited facts; illustrative examples are labelled.

Real organization, 2013-01

Sales incentives and mis-selling in UK retail finance (2013)

In January 2013 the Financial Services Authority, the predecessor of the Financial Conduct Authority, published guidance after reviewing incentive schemes at 22 banks, building societies, insurers and investment firms. It found that 20 of the 22 had features that increased the risk of mis-selling, significantly so in six cases, and that 11 of those 20 were not properly addressing the risk. Sales quality generally had much less effect on pay than the quantity sold; some schemes were so complex that management did not understand them; and some sales managers earned bonuses on the sales of the very staff whose sales they were meant to check. One scheme multiplied bonuses by up to eight times for cross-selling protection insurance once a target was passed, backdated over the whole month; another paid up to £10,000 to the first 21 people to reach a target.

Interpretation. This is the multitasking problem in its purest form: strong pay for measured volume, almost none for unmeasured suitability, and cliff edges where the temptation to cut corners peaks.

Real organization, 2018-03

TSB removes sales targets, as described by its chief executive (2018)

In an essay in the FCA's 2018 discussion paper on culture, TSB's chief executive and HR director wrote that the bank had scrapped all sales targets, sales-linked rewards and access to comparative sales data at branch and area director level, and rewarded staff on the service they gave customers. Taking the John Lewis Partnership as inspiration, it replaced an inherited bonus structure with an annual profit share for all staff, worth around 10% of salary and awarded on service. Individuals were assessed on skill, attitude and behaviour rather than outputs, and the size of the pool depended on fair customer outcomes, Net Promoter Scores and complaints resolution. They reported that the bank's Net Promoter Score had risen from −24 at its 2013 launch to +25 in 2017.

Interpretation. These are the company's own claims, not an independent evaluation. The design follows the multitasking logic: where quality cannot be measured person by person, weaken individual output pay and reward broad, shared measures.

Real organization, 2023-10

Removing the bankers' bonus cap (2023)

In October 2023 the FCA and the Prudential Regulation Authority removed the limit on the ratio of variable to fixed pay, the bonus cap, for banks, building societies and PRA-designated investment firms, from 31 October 2023. The FCA said the cap had led to growth in fixed pay, reducing firms' ability to adjust variable pay to absorb losses or to respond to poor performance or misconduct discovered later. The remuneration regime continued to rely on linking awards to risk-adjusted performance, payment in instruments, deferral, and malus and clawback.

Interpretation. A cap limits the intensity of incentives, but it also turns pay into salary, which cannot be clawed back. The regulators preferred a larger pool of adjustable, deferred pay to a smaller pool of capped bonuses.

Section 05: Case studies

Fictional case, simulated figures

Fernhill Hospitals: a bonus, or better rotas?

Fernhill Hospitals, a fictional group of three hospitals, cannot keep its ward nurses. Vacancies are filled by agency staff at high cost, and experienced nurses spend much of their time inducting newcomers who then leave. The finance director proposes a retention bonus of £3,000 paid to every nurse on reaching two years' service. The chief nurse proposes instead to let ward teams set their own rotas within agreed staffing levels and to guarantee flexible shift patterns for nurses with caring responsibilities, at an estimated cost of £400,000 a year in extra scheduling and some overlap of shifts. The board has the figures below.

Fernhill Hospitals ward nursing workforce, last 12 months
Measure Value
Funded nurse posts1,200
Average nurses in post1,080
Nurses who left during the year216
Leavers in their first two years of service120
Nurses reaching two years' service in a typical year150
Estimated cost of replacing one nurse£14,000
Agency spending to cover vacant shifts£9.0m
Exit survey, main reason: shift patterns and lack of flexibility74 (41%)
Exit survey, main reason: workload and staffing45 (25%)
Exit survey, main reason: pay40 (22%)
Exit survey, main reason: career development21 (12%)

Simulated figures. 180 of the 216 leavers answered the exit survey, each giving one main reason. Replacement cost covers recruitment, induction and lost output.

Questions

  1. Calculate the vacancy rate, the labour turnover rate, the share of leavers who go within two years and the annual cost of replacing leavers.
  2. What would the bonus cost a year at current numbers, and how many fewer leavers would it need to pay for itself on replacement costs alone? Why would its cost rise if it worked?
  3. Using the exit survey and the theories in this module, compare the bonus with self-rostering. What might each sort, and what might each crowd out?
  4. What would a make-or-buy view of the nursing workforce suggest about the agency spending?

Section 06: Critical evaluation

What the evidence supports, where the debate is, and what the models leave out.

What is established. Incentives change behaviour, often powerfully, when output is measurable and the job is simple: the Safelite evidence is typical of that literature. Sorting effects are frequently as large as effort effects. The side effects of incentives on unmeasured tasks are also well documented, in the economics of multitasking, in studies of timing games around thresholds and in regulators' reviews of mis-selling.

What is debated. How much crowding out matters in ordinary employment is disputed: laboratory meta-analyses disagree about its size, and the field studies are striking but specific. Popular books that claim bonuses always damage creativity generalize far beyond the evidence. Goal setting has a robust main effect; its dark side is documented mainly through cases and experiments.

Talent claims. The war for talent rested largely on consultants' surveys. The evidence that stars lose performance when they move comes mostly from equity analysts and may not generalize. Evidence on forced ranking in real firms is thin and confounded by everything else that changes alongside it.

A caution on behavioural fixes. A widely cited study reporting that people cheated less when they signed an honesty statement at the top of a form rather than the bottom (Shu and colleagues, 2012) was retracted in 2021 after its field data were found to be unreliable. Small, cheap interventions that promise large effects on employee behaviour should be tested locally before they are adopted at scale.

What the models leave out. Most incentive models treat pay as private and individual. In practice people compare their pay with colleagues', so pay structures raise questions of fairness, which equity theory addresses, and of pay transparency, which changes those comparisons.

Section 07: Decision exercise

A fictional situation. Decide before you read the guidance.

Redesigning an adviser's pay

Mason & Pell, a fictional independent insurance broker, pays its 60 advisers a modest salary plus commission averaging 40% of their pay, with an accelerator once a quarterly target is passed. Revenue has grown strongly, but complaints have doubled in two years and file checks find that one policy in ten sold near quarter-end lacks evidence that it suits the customer. The regulator has asked to see the incentive scheme.

  1. A. Keep commission, add a quality gate

    Keep commission and the accelerator, but withhold commission on any sale that fails a file check, and claw back pay on policies cancelled within a year.

  2. B. Salary and a shared profit pool

    Raise salaries, remove individual commission, and pay an annual team profit share scaled by complaint rates, file-check results and customer retention.

  3. C. Salary only, with promotion on quality

    Pay salary only and base promotion on peer review, customer feedback and file-check results over two years.

Judge the options on

  • Effort and revenue
  • Customer outcomes and conduct risk
  • Which advisers the scheme attracts and keeps
  • Cost and reliability of measuring quality
  • The regulator's view
Read the guidance

The evidence points away from the status quo: the scheme pays strongly for volume, weakly for suitability, and its accelerator creates the quarter-end cliff where the file checks find problems. A keeps that structure and relies on the file checks to catch what the commission encourages. It works only if checks are frequent and reliable; otherwise advisers learn what passes a check, and the gate becomes one more target to manage.

B fits the multitasking logic best. It weakens individual output pay where quality is hard to measure and ties a shared reward to broad outcomes. Expect sorting: some high-volume advisers will leave. That cost is real, but if their volume came with conduct risk, part of it was never profit. C removes the distortion entirely but leans heavily on culture and on the credibility of promotion decisions, and it may lose effort as well as risk.

A defensible answer is B, phased in over a year with the accelerator removed at once, the profit pool measures published in advance, and a transition payment to avoid a sudden cut in pay for advisers with good file records.

Section 08: Further reading

  • Steven Kerr (1975). On the folly of rewarding A, while hoping for B. Academy of Management Journal, 18(4).

    The classic, readable diagnosis of reward systems that pay for one thing while hoping for another.

  • Robert Gibbons (1998). Incentives in organizations. Journal of Economic Perspectives, 12(4).

    A clear economist's survey of incentives in organizations, from agency theory to subjective performance evaluation.

  • Bengt Holmström and Paul Milgrom (1991). Multitask principal-agent analyses: incentive contracts, asset ownership, and job design. Journal of Law, Economics, and Organization, 7 (special issue).

    The multitask model. Technical, but the introduction and examples are accessible.

  • Edward P. Lazear (2000). Performance pay and productivity. American Economic Review, 90(5).

    The Safelite study; a model of how to separate incentive effects from sorting.

  • Edward L. Deci, Richard Koestner and Richard M. Ryan (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6).

    The most cited meta-analysis on rewards and intrinsic motivation; read with Cameron and Pierce (1994) for the other side.

  • Peter Cappelli (2008). Talent on Demand: Managing Talent in an Age of Uncertainty. Harvard Business Press.

    Talent management as a supply chain problem, with practical alternatives to rigid succession planning.

  • Boris Groysberg, Linda-Eling Lee and Ashish Nanda (2008). Can they take it with them? The portability of star knowledge workers' performance. Management Science, 54(7).

    Evidence on why star performers often fail to repeat their success after moving.

Every work cited here is listed in the platform's sources register. Real organizations appear only with facts from the sources shown; everything else is labelled illustrative or fictional.